Low certainty * "large reduction in risk of infection" doesn't equal masks being marginal though, does it? (Not rhetorical, I'm an amateur at statistics) That means they need more data, but if an effect exists, it is large?

And I feel like we can reason about this to some extent from first principles - If the virus is spread through particles from faces, then face coverings will decrease the numbers of particles from faces that 1) get into the air and 2) get into peoples' faces, no matter how crappy the face coverings are. It seems more plausible to me that the problem is lack of data rather than incorrect conclusion, which seems like a reasonable interpretation of this particular study?